6

I have a quite large table (45 columns x 15 rows). Columns include comparisons groups and each row represents a measure. I can only think about putting the table on multiple pages but this does not make it easy to compare groups' estimates that are on different pages? What are good practices for such cases?

Option 1: multi-page table?

Table 1, Table 1 (continued 1), Table 1 (continued 2) ...

What are other good options for formatting large tables?

4
  • 1
    I was once told the following, which is almost always true "Don't say anything in a table, that could have been said in a figure". Hence, think about other ways you could convey the message from the table.
    – Sune
    Jun 21 at 17:23
  • I'll note that good options for formatting larger tables depend heavily on the type of data and the level of contextual interdependency between the columns (i.e., whether understanding a given column requires context from other columns). Assuming you do want/need to retain this table (the accepted answer hints otherwise), we don't have enough information to offer advice on how to accomplish this.
    – Brian
    Jun 21 at 17:32
  • In addition to other suggestions, transposing the table is likely to be better, if at all possible. Even if you're not using LaTeX, you may find some good suggestions at Te.X.se; here's a search for [tables]+multipage to give you an idea
    – Chris H
    Jun 22 at 12:06
  • To those questioning the validity of producing such a table in the first place: yes, it's probably not a good idea to try and use a gigantic table to highlight the key trends in the data. However, it's often very helpful to include a comprehensive table in an appendix. There are a number of papers where I often refer to (appendix) tables for specific measurements of specific objects. Yes, eventually a point is reached where going machine readable is the only sane choice, but in some contexts a table spanning a few pages is a very helpful thing.
    – Kyle
    Jun 22 at 22:24
1

I assume you seek to include a large table in print thesis and you live in a country with access to paper in either one of ISO A series (if unfamiliar to this definition, see here) and you already know which data really need to stay (ask a colleague, your supervisor).

Do not shrink the font size to add more content on an already jammed page, but consider to print the large table on a sheet of ISO A3 (landscape orientation), fold it, and bind this into the rest of your thesis print on ISO A4 (portrait orientation). This is much easier (than with standard US paper formats)* because within the ISO A series, the long side of a smaller paper format is as long as the short side of the next larger paper format.

You do not need to fill all the A3 page in the horizontal direction. It actually is better to use the same font size on the lager page as in the other part of your thesis. For the potential trim to a smaller size of «the table page» (then neither ISO A3, nor ISO A4) and folding, reach out for help by the staff members of a good printer's shop, who will bind it with all the other pages into your thesis for you.

It equally is a technique you find for business reports comparing selected key figures of debit and credit about the year to be reported side-by-side with the numbers about year before the one of principal interest. On occasion, architects, engineers, etc. use it to include their plans and drawings. For illustration, see for example this or even this video.

*) It is not an insurmountable obstacle for a good printer's shop.

26

There is no need to reproduce your entire raw data in your thesis (or any other publication). Your publication should describe some abstract properties of your data, discuss your analysis, and present your results. As a rule, if a table spans more than two adjacent pages, it is too large. Ask yourself: What is the point I want to get across by presenting the data in the chosen format? What do I need to focus on to get this point across more efficiently and without distraction?

If you want to make your data accessible, which at some point could be a good idea, consider uploading it to a research repository or publishing it in an (online) appendix.

Don't forget to ask your supervisor. Whatever they say obviously overrides any advice you may get from strangers on the internet.

3
  • 6
    There are thesis committees all over the world which disagree with you on this. ;-) It might not be ideal, but it’s often required of students to include extensive, pages-spanning tables into the appendix of the thesis document. Having separate files is not generally acceptable. Jun 21 at 16:55
  • 1
    Yes, it's best to ask the supervisor, but the issue looks like there is a deeper problem underneath that's about analysis and presenting the argument rather than layout per se.
    – henning
    Jun 21 at 17:01
  • 2
    The key word in @KonradRudolph's comment, IMO, is appendix. In my case it was code, but in a similar way you should present enough to support your conclusions, in a suitable format (often a figure), in the body text; if there are reasons (diktats perhaps) to keep the bulk data with the publication, then an appendix (or, for a paper, supplementary information) is best. There are cases in which tabular data is an important output, but they're rare these days.
    – Chris H
    Jun 22 at 12:03
5

@henning answer is absolutely on point, to add just another point. Please highlight what is the objective of this data, what kind of difference it shows and represent it in values rather than publishing the raw table.

In any case this video about data visualization might help you to get your point through. But in nutshell each table has a story to show where the objective is relativity to the other values. Thus highlight the relation and not the data.

Storytellingwithdata-googleTalks

2

@henning's answer already mentioned getting your advisor's input, but before that, I would suggest checking your university's thesis format requirements. My university, for example, has guidelines specifically for multi-page tables (I had to know this for what ended up being a 15 page table in my masters thesis). Then, within whatever guidelines your university may or may not have on the matter, you can go to your advisor and get their input. You don't want to assume your advisor is aware of every little formatting requirement your university requires.

1

I do not understand what is in the table, but if there are 45 important columns and 15 rows, one can consider to make each row a full page where you list the column items, maybe add an executive summary per this page and then make a new table where only 15 rows and 2-5 columns (the gist) is displayed.

This is if all the information is really important and cannot be transformed to graphs.

0

@henning's answer has a great point about whether or not you need to show all your data, but you might want to think about whether you need to show all the data in a single table. If you have multiple points to make it might be more useful to break the one giant table into multiple smaller tables that each have a single focus.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.